I gave an AI agent three real-world cybersecurity consulting challenges to find out whether it could actually do my job.

The results surprised me.

In this episode of AI Security Multipliers, I put ChatGPT to work as a senior cybersecurity consultant for a fictional global manufacturing company with 12,000 employees, legacy technology, cybersecurity risks, budget constraints and executives with competing priorities.

I gave it three assignments:

1. Conduct a cybersecurity risk assessment.
2. Create an executive briefing for business leadership.
3. Navigate a difficult decision involving security, cost, production and AI adoption.

The agent delivered some impressive results. It identified meaningful cybersecurity risks, produced practical recommendations and demonstrated surprisingly good business reasoning.

But when I asked it to create a professional executive PowerPoint, things got interesting. Even after feedback, the presentation still wasn't something I'd put in front of a client.

We scored the three challenges at 78, 68 and 85 out of 100, with an overall average of 77. These were subjective assessments of the deliverables, not a formal benchmark.

So, can an AI agent replace an experienced cybersecurity consultant?

In this episode, I share what worked, what disappointed me, where AI agents still need human judgment and whether I'd trust one to work independently.

The Multiplier: A practical experiment you can run yourself to determine whether an AI agent genuinely saves time and delivers work that meets your standards.

Because there's a big difference between an AI agent completing a task and delivering something you'd be willing to put your name on.

AI Security Multipliers explores the intersection of artificial intelligence, cybersecurity, risk and governance, with practical insights for technology leaders and security professionals.

Send us Fan Mail